The global burden of tuberculosis (TB) has increased over the past two decades despite widespread implementation of control measures, suggesting that many of these strategies could be improved. Contact patterns of an index case and the susceptibility of the exposed individual are critical determinants of TB transmission. Contact patterns are significantly influenced by the existence of networks within a population. At the same time, HIV-infection has been the most important factor in the resurgence of TB worldwide by increasing the susceptibility of the exposed individual. However, no studies have carefully looked into the effect of transmission networks and HIV-infection in the dynamics of TB epidemics in the community. Most of the current infection control strategies are based on the concept that TB acquisition requires prolonged and close contact with an infectious case. However, recent TB outbreak investigations, affecting primarily HIV-infected patients, have failed to identify any link or evidence of close contact among cases infected with the same strain of M. tuberculosis, challenging this widely spread concept. Contrary to the establish paradigm, we believe that casual encounters, occurring within unrecognized transmission networks, are important drivers of TB epidemics in dynamic populations with high prevalence of HIV infection. Botswana offers unique opportunities to determine the role of HIV, the degree of contact to an index case and unrecognized social networks in TB epidemics. Over the last 6 years, we have developed a nationwide TB surveillance program in Botswana. We have seen that new TB cases without clear exposure to an index case are more frequently seen in the Southern areas of the country and are more frequently HIV- associated. Consistent with this observation, molecular genotyping of M. tuberculosis from sputum samples from HIV-infected patients from this area show that, despite the absence of known contact with an index case, most of them are infected with the same strain, suggesting recent infection from a common source. In contrast, TB in the Western region, which has one of the highest rates of TB in the world, is more HIV-independent, and close contact with an index case is often evident. Thus, we believe that there are distinct epidemics within the larger TB epidemic in Botswana that are fueled by different mechanisms of transmission and acquisition, probably related to the population dynamics (close vs. casual exposures) and prevalence of HIV infection. In this proposal, we aim to determine the relative role of these factors (HIV, the degree of contact to an index case and unrecognized social networks) as drivers of the TB epidemics in these 2 populations by combining classical epidemiologic methods and molecular epidemiology. We expect this study will show that casual encounters, primarily occurring in unrecognized transmission networks, are the main drivers of TB epidemics in populations with high prevalence of HIV, which would represent a change in current paradigms.